Measuring change from alcohol brief interventions: development of an international consensus core outcome set

3 January 2023


Gillian W Shorter 1,2, Nick Heather3, Jeremy W Bray4, Amy J O’Donnell5, Aisha Holloway6, Emma L Giles2, Anne H Berman7,8, Mike Clarke1, Carolina Barbosa9, & Dorothy Newbury-Birch2

1Queen’s University Belfast, UK; 2Teesside University, UK; 3Northumbria University, UK; 4University of North Carolina at Greensboro, USA; 5Newcastle University, UK; 6Edinburgh University, UK; 7Karolinska Institutet, Sweden; 8Uppsala University, Sweden; 9RTI International, USA.

Key findings

10 items should be reported in all efficacy and effectiveness trials of Alcohol Brief Interventions (ABI) in adults aged 16 or over who are not seeking formal treatment.

  • Typical frequency of consumption
  • Typical quantity of consumption
  • Frequency of heavy episodic drinking
  • A combined consumption measure, which summarises alcohol use
  • Hazardous or harmful drinking
  • Standard drinks consumed in a week
  • Alcohol related consequences
  • Alcohol related injury
  • Use of emergency healthcare services
  • Quality of life

These should be measured to:

  1. Reduce selective reporting (e.g. only reporting some, often significant findings)
  2. Support stronger reviews of ABI evidence, as more studies can be included
  3. Reduce research waste (e.g. so your work can be used by more researchers)
  4. Help researchers easily find the internationally important outcomes
  5. Minimise time taken to justify outcomes to funders.

Reported outcomes should include, in detail:

  • What was measured, how it was measured (including statistical methods and time points), and effect sizes with confidence intervals.

Trial reports should also comply with the Consolidated Standards for Reporting Trials (CONSORT), and relevant extensions.


Alcohol brief interventions (ABIs) are used to help address hazardous and harmful alcohol use in a range of settings for adults not seeking formal treatment. However, assessing if ABIs work is difficult because reported outcomes differ by trial, and are measured in many different ways (e.g. Kaner et al., 2018; Khadjesari et al., 2011). The differences in what we measure and how we report it makes producing statistical summaries of evidence (meta-analyses) very difficult. It increases selective reporting where not all outcomes are reported or research waste where papers cannot be used by the wider field to summarise if an intervention works (Glasziou, 2014).

‘Core Outcome Sets’ are the minimum outcomes that should be measured and reported in all trials (Williamson et al., 2017). This has improved evaluation quality and evidence synthesis in other healthcare settings and helped researchers justify outcome choice as the international standard. We aimed to develop a Core Outcome Set for ABI trials which can improve evidence reviews and increase quality in evaluation and reporting. This will help us see which ABIs work and in what settings. Researchers can measure other outcomes with the core outcomes to support their project aims.


The project had three phases guided by a protocol (Shorter et al. (2017))

  1. Systematic review to show what outcomes are used and how these are measured
  2. E-Delphi prioritisation exercise to rank the importance of these outcomes
  3. Consensus meeting of experts to agree the final Core Outcome Set


Systematic review

The systematic review identified 2641 outcomes from 401 peer-reviewed papers, measured in around 1560 different ways (Shorter et al., 2019a). The average trial had 450 participants. Most trials had two arms, and 83% had a non-ABI control group. There were more efficacy than effectiveness trials. On average there were two primary, and four secondary outcomes. Compliance with basic elements of trial reporting guidance (CONSORT) was low, with around half including “trial” in the title, and around 60% had a participant flow chart. Outcomes differed by setting, with biomarkers the least frequent of outcomes. The most common outcomes were those measuring consumption.

e-Delphi survey

A two-round e-Delphi survey (Shorter et al., 2019b) showed preferred outcomes from researchers, healthcare professionals, and people with lived/living experience from 19 countries. Outcomes ranked as important by 60%+ participants went forward to the consensus meeting. Fifteen outcomes met the criteria:

  • Typical drinking frequency
  • Frequency of heavy drinking
  • Number of standard drinks in a week
  • Hazardous or harmful drinking
  • Alcohol related problems
  • A combined consumption measure
  • Typical quantity
  • Readiness to change by the participant
  • Alcohol related injury
  • Use of alcohol or drug treatment
  • Psychological wellbeing
  • Whether the intervention was delivered as planned/participant used intervention
  • Emergency healthcare
  • Hospitalisation
  • Quality of life

Consensus meeting

Sixteen delegates from seven countries and a range of ABI stakeholder groups, attended a full day consensus meeting in New York.

  • Each outcome from the e-Delphi above was discussed, and a vote was cast
  • If a majority supported inclusion of the outcome, it was added to the set.
  • After all outcomes had been voted on individually, the full Core Outcome Set was discussed to reduce redundancy or overlap between the outcomes.
  • A vote was cast for the final set.
  • Potential measures from the systematic review were reviewed for their psychometric properties and voted on by consensus meeting delegates and individuals with lived/living experience of unhealthy alcohol use.

The ten items and their measures are freely available see Shorter et al., (2021) for a data dictionary and the Open Science Framework page for details on decision making.

Implications and conclusion

We should measure the ORBITAL core outcome set in all efficacy and effectiveness trials of ABIs that target people at risk of or currently experiencing alcohol related harm, who are over the age of 16, and who are not seeking formal alcohol treatment. This will improve trial reporting, ensure we are measuring the outcomes that matter to stakeholders internationally, help researchers justify their outcome choice, reduce selective reporting, and ensure evidence synthesis can include more evidence. Other outcomes can be measured alongside the set which is the minimum.

The outcomes include average consumption measures which summarise use over time

  1. Typical frequency of consumption
  2. Typical quantity of consumption
  3. Frequency of heavy episodic drinking
  4. Combined consumption measure
  5. Hazardous or harmful drinking1

Recent consumption reflecting current alcohol use

6. Standard drinks consumed in the past week (in grams)2

Impact of alcohol use summarising the key negative effects of alcohol use

7. Alcohol-related consequences3
8. Alcohol-related injury4
9. Use of emergency health care services for any reason5

Quality of life summarising the standard of health, comfort, or happiness of the drinker

10. Quality of life6


1 Measured using AUDIT-C (Bush et al., 1998). Include a guide to a standard drink in presentation.
2 Adapted from Sobell & Sobell (1992) – convert standard drinks per day into a summary number of grams for the week.
3 Short Index of Problems (Feinn et al., 2003)
4 Modified/adapted question from the Short Index of Problems Feinn et al. (2003)
5 Modified/adapted question from Econ Form-90 Bray et al. (2007)
6 Either PROMIS Global Health 1.2 or WHOQOL-Bref Hays et al. (2009); WHOQOL group (1998)